3.5mm headset to XLR adaptors please

[kozikowski]

I think I know of one reason Sweetwater isn't going to be interested in this. We're not actually converting a gamer microphone into an XLR-3 Male connection. We're only faking it poorly so that a very short XLR cable can get the performance into the mixer before it starts buzzing and humming.

A real XLR-3 connection such as from a Rock Band Microphone will work with a hundred foot long cable (attached picture). This adapter won't do that. So the first time a singer tries to use this adapter in a real performance, or you try to use it with long cables, everybody is going to be in deep trouble.

Koz

[flynwill]

Back on the subject of using the Phantom power....

The IEC phantom power standards says you can draw up to 10 ma.

The standard computer mic connection is 5V through 2.2k meaning the most you can draw is about 20 ma (at 0 volts), but if you want some useful voltage you'd have to be under 10ma.

So all you need is a mechanism to drop the voltage. I would suggest a 4.2 - 10k resistor (for 10 ma or 4 ma max respectively) and a 6V Zener.

However, I agree that at some point it's better to just go buy a real microphone.

[kozikowski]

That would be a noisy zener in a microphone circuit? That's the other problem with working signals in the .001 volt range. That and we still haven't created a real balanced XLR signal.

Those AKG microphones seem to work pretty well.

Koz

[flynwill]

Well yes you need a bypass cap on the zener. I doubt it would be any noisier than what is provided by the typical PC. If its a real concern then I'm back to "buy a real microphone".

[kozikowski]

I wonder what you would need to do this job for real. Say one of the requirements is the mixer has to supply 48v on pins 2 and 3. Why would you even need the zener? Just keep increasing the size of the load resistor so the "plate current" remains the same. How would you create a balanced signal on 2 and 3?

I know there's a trick to this. You don't have to drive both sides, but the impedances have to match.....

Koz

[flynwill]

Well to create a truly balanced signal is going to either need active components or a transformer. The scheme I was envisioning pin 3 would have no AC signal relative to pin 1, and all of he signal would be in pin 2. (There would probably be a fat capacitor between 3 and 1).

A single transistor inverting gain one amp could get you the inverted signal, and easily be powered from the same 48V. However to match the impedances you'd probably end up adding a second transistor as an emitter follower on the mic input so that you could control it's impedance. (Because you have no idea what the output impedance of the computer microphone is.)

The zener is there for the same reason. You don't know how much current the computer mic needs, and I would presume that 48V would fry it should it only need 10-20 uA instead of 10 mA. OTOH I just realized that a simple resistor divider will work..... The typical circuit in the computer is 2.2K to 5V. A 22k resistor and 2.5k resistor wired as a divider will get you the same thing effectively.

[kozikowski]

to create a truly balanced signal is going to either need active components or a transformer.

I went to a learned lecture once which refuted that. Almost certainly if you needed a floating sound generator you'll be pricing transformers — either to generate the signal or to manage the DC, but to get a balanced generator requires equal resistances in the signal and ground legs. Full stop.

Thinking about that is giving me a headache, but I wonder if that can be put to good use. Yes, the output impedance of the system without the zener is going to be roughly the value of the "plate load resistor," so running the system directly on 48V is not useful. But instead of jamming pins 1 and 3 together (everyone's natural reaction) build out pin 3 so it matches pin 2 in resistance if not signal.

Line receivers are looking for a difference between 2 and 3 to derive the show. Any difference. So far, so good. Anything appearing on both 2 and 3 is rejected. If the branch resistance on both legs is identical, radiation or other interference on both wires will match.

I think the only thing I'm missing is a "good" microphone's output impedance of 150 ohms.

That would be rough to do without some very serious work.

Do the DC to DC converters "act like" a transformer in that the current availability goes up when the voltage goes down?

Koz

[flynwill]

So I think this will work:

XLRAdapt1.jpg (20.65 KiB) Viewed 1961 times

It is definitely not balanced and probably won't run down a long cable without noise, loss or other issues. This design will definitely "pop" when you plug it in (or if you turn the phantom power on/off) so you want to be sure the volume down when connecting.

If you want a balanced output then you need to add a matching transformer like this:

XLRAdapt2.jpg (20.42 KiB) Viewed 1961 times

But given a decent transformer is a pretty good fraction of the price of a decent microphone why bother?

PS: The caption on the J2 should read "mini-phone" not "mini-phono"...

PPS: On reading the phantom power spec... 48V power is supposed to have 6.8k series resistance from the source, so in the first design the 22k R1 should probably be 15k. In the second design (since it is effective fed by both legs) R1 should be 18k.

[flynwill]

A little web searching turned up this circuit:

http://www.twin-x.com/groupdiy/albums/u ... it_Mic.gif

Which with a bit of modification would probably work as well.

I also found a significantly more complicated circuit here:

http://electronicdesign.com/site-files/ ... fig_01.gif
and accompanying explanation here: http://electronicdesign.com/power/modif ... distortion

which is a lot more components, but relatively cheap ones.

Note both of these designs use zener diodes to establish their V+ for the amp circuit, and both use a FET source-follower to create the + and - signals.

[kozikowski]

So far no store-bought devices?
Koz